ALBERT

Description: We present a 2-D subsurface image of the Paganica Fault from a high-resolution refraction tomography and detailed geological investigation carried out across part of the northwestern segment of the 20-km-long Paganica–San Demetrio fault-system, and which was responsible of the 2009 April 6 M w 6.1 L'Aquila earthquake (central Italy). We acquired two seismic profiles crossing the Paganica basin with a dense-wide aperture configuration. More than 30 000 P wave first-arrival traveltimes were input to a non-linear tomographic inversion. The obtained 250–300 m deep 2-D Vp images illuminate the shallow portion of the Paganica Fault, and depict additional unreported splays defining a complex half-graben structure. We interpret local thickening of low- Vp (〈2400 m s –1 ) and intermediate- Vp (2600–3400 m s –1 ) regions as syn-tectonic clastic wedges above a high- Vp (3800–5000 m s –1 ) carbonate basement. These results are condensed in a 4.2-km-long section across the Paganica basin, clearly indicating that the Paganica Fault is a mature normal fault cutting the whole upper ~10 km of the crust. We evaluate a minimum cumulative net displacement of 650 ± 90 m and a total heave of 530 ± 65 m accomplished by the Paganica Fault, respectively. In the conservative hypothesis that the extension started during the Gelasian (1.80–2.59 Ma), we obtain a minimum long-term slip-rate of 0.30 ± 0.07 mm yr –1 and an extension-rate of 0.25 ± 0.06 mm yr –1 , respectively. Considering the regional averaged extensional field of ~1 mm yr –1 obtained from geodetic and geological analyses at 10 4  yr timescale, we infer that the Paganica Fault accounts for ~20 per cent of the NE-extension affecting this zone of the central Apennines axis due to the concurrent activity of other parallel normal fault-systems nearby (e.g. the Liri, Velino-Magnola, L'Aquila-Celano and Gran Sasso fault-systems).

Description: We use recently published, high-resolution reconstructions of the Southwest Indian Ridge to test whether a previously described systematic difference between Global Positioning System (GPS) and 3.16-Myr-average estimates of seafloor spreading rates between Antarctica and Africa is evidence for a recent slowdown in Southwest Indian Ridge seafloor spreading rates. Along the Nubia-Antarctic segment of the ridge, seafloor opening rates that are estimated with the new, high-resolution reconstructions and corrected for outward displacement agree well with geodetic rate estimates and reduce previously reported, highly significant non-closure of the Nubia-Antarctic-Sur plate circuit. The observations are inconsistent with a slowdown in spreading rates and instead indicate that Nubia-Antarctic plate motion has been steady since at least 5.2 Ma. Lwandle-Antarctic seafloor spreading rates that are estimated from the new high-resolution reconstructions differ insignificantly from a GPS estimate, thereby implying steady Lwandle-Antarctic plate motion since 5.2 Ma. Between the Somalia and Antarctic plates, the new Southwest Indian Ridge reconstructions eliminate roughly half of the systematic difference between the GPS and MORVEL spreading rate estimates.We interpret the available observations as evidence that Somalia-Antarctic spreading rates have been steady since at least 5.2 Ma and postulate that the remaining difference is attributable to random and/or systematic errors in the plate kinematic estimates and the combined effects of insufficient geodetic sampling of undeforming areas of the Somalia plate, glacial isostatic adjustment in Antarctica and transient deformation triggered by the 1998 M w = 8.2 Antarctic earthquake, the 2004 M w = 9.3 Sumatra earthquake, or possibly other large historic earthquakes.

Description: We present a new, computationally efficient numerical method to simulate global seismic wave propagation in realistic 3-D Earth models. We characterize the azimuthal dependence of 3-D wavefields in terms of Fourier series, such that the 3-D equations of motion reduce to an algebraic system of coupled 2-D meridian equations, which is then solved by a 2-D spectral element method (SEM). Computational efficiency of such a hybrid method stems from lateral smoothness of 3-D Earth models and axial singularity of seismic point sources, which jointly confine the Fourier modes of wavefields to a few lower orders. We show novel benchmarks for global wave solutions in 3-D structures between our method and an independent, fully discretized 3-D SEM with remarkable agreement. Performance comparisons are carried out on three state-of-the-art tomography models, with seismic period ranging from 34 s down to 11 s. It turns out that our method has run up to two orders of magnitude faster than the 3-D SEM, featured by a computational advantage expanding with seismic frequency.

Description: Tehran, the capital city of Iran with a population of over 12 million, is one of the largest urban centres within the seismically active Alpine–Himalayan orogenic belt. Although several historic earthquakes have affected Tehran, their relation to individual faults is ambiguous for most. This ambiguity is partly due to a lack of knowledge about the locations, geometries and seismic potential of structures that have been obscured by dramatic urban growth over the past three decades, and which have covered most of the young geomorphic markers and natural exposures. Here we use aerial photographs from 1956, combined with an ~1 m DEM derived from stereo Pleiades satellite imagery to investigate the geomorphology of a growing anticline above a thrust fault—the Pardisan thrust—within central Tehran. The topography across the ridge is consistent with a steep ramp extending from close to the surface to a depth of ~2 km, where it presumably connects with a shallow-dipping detachment. No primary fault is visible at the surface, and it is possible that the faulting dissipates in the near surface as distributed shearing. We use optically stimulated luminescence to date remnants of uplifted and warped alluvial deposits that are offset vertically across the Pardisan fault, providing minimum uplift and slip-rates of at least 1 mm yr –1 . Our study shows that the faults within the Tehran urban region have relatively rapid rates of slip, are important in the regional tectonics, and have a great impact on earthquake hazard assessment of the city and surrounding region.

Description: We report on a study to explore the deep electrical conductivity structure of the Dead Sea Basin (DSB) using magnetotelluric (MT) data collected along a transect across the DSB where the left lateral strike-slip Dead Sea transform (DST) fault splits into two fault strands forming one of the largest pull-apart basins of the world. A very pronounced feature of our 2-D inversion model is a deep, subvertical conductive zone beneath the DSB. The conductor extends through the entire crust and is sandwiched between highly resistive structures associated with Precambrian rocks of the basin flanks. The high electrical conductivity could be attributed to fluids released by dehydration of the uppermost mantle beneath the DSB, possibly in combination with fluids released by mid- to low-grade metamorphism in the lower crust and generation of hydrous minerals in the middle crust through retrograde metamorphism. Similar high conductivity zones associated with fluids have been reported from other large fault systems. The presence of fluids and hydrous minerals in the middle and lower crust could explain the required low friction coefficient of the DST along the eastern boundary of the DSB and the high subsidence rate of basin sediments. 3-D inversion models confirm the existence of a subvertical high conductivity structure underneath the DSB but its expression is far less pronounced. Instead, the 3-D inversion model suggests a deepening of the conductive DSB sediments off-profile towards the south, reaching a maximum depth of approximately 12 km, which is consistent with other geophysical observations. At shallower levels, the 3-D inversion model reveals salt diapirism as an upwelling of highly resistive structures, localized underneath the Al-Lisan Peninsula. The 3-D model furthermore contains an E–W elongated conductive structure to the northeast of the DSB. More MT data with better spatial coverage are required, however, to fully constrain the robustness of the above-mentioned off-profile features.

Description: Mid-crustal conductors are a common phenomenon in magnetotelluric studies. In the Andean Cordillera of southern Chile, they appear to concentrate along major fault zones. A high-resolution, broad-band magnetotelluric survey including 31 stations has been carried out along two profiles perpendicular to (1) the Liquiñe-Ofqui Fault Systems (LOFS) and (2) the Villarrica-Quetrupillán-Lanín volcanic lineament running parallel to the Mocha-Villarrica Fault Zone (MVFZ). The survey aimed at tracing one of the known conductors from mid-crustal depth to near-surface along these faults. Directionality and dimensionality were analysed using tensor decomposition. Phase tensors and induction arrows reveal two major geoelectric strike directions following the strike of LOFS and MVFZ. 2-D inversion shows low resistivity zones along both fault systems down to a depth of 〉10 km, where the brittle-ductile transition is expected. Along the LOFS, the two anomalies are linked to (1) Lake Caburgua, where the LOFS broadens to about 2 km of lateral extension and seems to represent a pull-apart structure, and (2) the intersection with the Villarrica-Quetrupillán-Lanín volcanic lineament, where seismic activity was observed during the latest eruption in March 2015. A connection of the mid-crustal conductor to the ESE-WNW-striking fault zones is indicated from the presented data.

Description: Melt generation and migration are an important link between surface processes and the thermal and chemical evolution of the Earth's interior. However, their vastly different timescales make it difficult to study mantle convection and melt migration in a unified framework, especially for 3-D global models. And although experiments suggest an increase in melt volume of up to 20 per cent from the depth of melt generation to the surface, previous computations have neglected the individual compressibilities of the solid and the fluid phase. Here, we describe our extension of the finite element mantle convection code ASPECT that adds melt generation and migration. We use the original compressible formulation of the McKenzie equations, augmented by an equation for the conservation of energy. Applying adaptive mesh refinement to this type of problems is particularly advantageous, as the resolution can be increased in areas where melt is present and viscosity gradients are high, whereas a lower resolution is sufficient in regions without melt. Together with a high-performance, massively parallel implementation, this allows for high-resolution, 3-D, compressible, global mantle convection simulations coupled with melt migration. We evaluate the functionality and potential of this method using a series of benchmarks and model setups, compare results of the compressible and incompressible formulation, and show the effectiveness of adaptive mesh refinement when applied to melt migration. Our model of magma dynamics provides a framework for modelling processes on different scales and investigating links between processes occurring in the deep mantle and melt generation and migration. This approach could prove particularly useful applied to modelling the generation of komatiites or other melts originating in greater depths. The implementation is available in the Open Source ASPECT repository.

Description: We present numerical models of mantle dynamics forced by plate velocities history in the last 450 Ma. The lower-mantle rheology and the thickness of a dense basal layer are systematically varied and several initial procedures are considered for each case. For some cases, the dependence on the mantle convection vigour is also examined. The resulting evolution of the CMB heat flux is analysed in terms of criteria to promote or inhibit reversals inferred from numerical dynamos. Most models present a rather dynamic lower mantle with the emergence of two thermochemical piles towards present-day. Only a small minority of models present two stationary piles over the last 450 Myr. At present-day, the composition field obtained in our models is found to correlate better with tomography than the temperature field. In addition, the temperature field immediately at the CMB (and thus the heat flux pattern) slightly differs from the average temperature field over the 100-km thick mantle layer above it. The evolution of the mean CMB heat flux or of the amplitude of heterogeneity seldom presents the expected correlation with the evolution of the palaeomagnetic reversal frequency suggesting these effects cannot explain the observations. In contrast, our analysis favours ‘inertial control’ on the geodynamo associated with polar cooling and in some cases break of Taylor columns in the outer core as sources of increased reversal frequency. Overall, the most likely candidates among our mantle dynamics models involve a viscosity increase in the mantle equal or smaller than 30: models with a discontinuous viscosity increase at the transition zone tend to agree better at present-day with observations of seismic tomography, but models with a gradual viscosity increase agree better with some of the criteria proposed to affect reversal frequency.

Description: Lateral viscosity variations (LVVs) in the mantle influence geodynamic processes and their surface expressions. With the observed long-wavelength geoid, free-air anomaly, gravity gradient in three directions and discrete, high-accuracy residual topography, we invert for depth- and temperature-dependent and tectonically regionalized mantle viscosity with a mantle flow model. The inversions suggest that long-wavelength gravitational and topographic signals are mainly controlled by the radial viscosity profile; the pre-Cambrian lithosphere viscosity is slightly (~ one order of magnitude) higher than that of oceanic and Phanerozoic lithosphere; plate margins are substantially weaker than plate interiors; and viscosity has only a weak apparent, dependence on temperature, suggesting either a balancing between factors or a smoothing of actual higher amplitude, but short wavelength, LVVs. The predicted large-scale lithospheric stress regime (compression or extension) is consistent with the world stress map (thrust or normal faulting). Both recent compiled high-accuracy residual topography and the predicted dynamic topography yield ~1 km amplitude long-wavelength dynamic topography, inconsistent with recent studies suggesting amplitudes of ~100 to ~500 m. Such studies use a constant, positive admittance (transfer function between topography and gravity), in contrast to the evidence which shows that the earth has a spatially and wavelength-dependent admittance, with large, negative admittances between ~4000 and ~10 4 km wavelengths.

Description: Three 2-D Deep Electrical Resistivity Tomography (ERT) transects, up to 6.36 km long, were obtained across the Paganica-San Demetrio Basin, bounded by the 2009 L'Aquila M w 6.1 normal-faulting earthquake causative fault (central Italy). The investigations allowed defining for the first time the shallow subsurface basin structure. The resistivity images, and their geological interpretation, show a dissected Mesozoic-Tertiary substratum buried under continental infill of mainly Quaternary age due to the long-term activity of the Paganica-San Demetrio normal faults system (PSDFS), ruling the most recent deformational phase. Our results indicate that the basin bottom deepens up to 600 m moving to the south, with the continental infill largely exceeding the known thickness of the Quaternary sequence. The causes of this increasing thickness can be: (1) the onset of the continental deposition in the southern sector took place before the Quaternary, (2) there was an early stage of the basin development driven by different fault systems that produced a depocentre in the southern sector not related to the present-day basin shape, or (3) the fault system slip rate in the southern sector was faster than in the northern sector. We were able to gain sights into the long-term PSDFS behaviour and evolution, by comparing throw rates at different timescales and discriminating the splays that lead deformation. Some fault splays exhibit large cumulative throws (〉300 m) in coincidence with large displacement of the continental deposits sequence (〉100 m), thus testifying a general persistence in time of their activity as leading splays of the fault system. We evaluate the long-term (3–2.5 Myr) cumulative and Quaternary throw rates of most of the leading splays to be 0.08–0.17 mm yr –1 , indicating a substantial stability of the faults activity. Among them, an individual leading fault splay extends from Paganica to San Demetrio ne’ Vestini as a result of a post-Early Pleistocene linkage of two smaller splays. This 15 km long fault splay can explain the Holocene surface ruptures observed to be larger than those occurred during the 2009 L'Aquila earthquake, such as revealed by palaeoseismological investigations. Finally, the architecture of the basin at depth suggests that the PSDFS can also rupture a longer structure at the surface, allowing earthquakes larger than M 6.5, besides rupturing only small sections, as it occurred in 2009.